Efficiently Estimating Erdos-Renyi Graphs with Node Differential Privacy

Authors: Jonathan Ullman, Adam Sealfon

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We give a simple, computationally efficient, and node-differentially-private algorithm for estimating the parameter of an Erd os-Rényi graph that is, estimating p in a G(n, p) with near-optimal accuracy. Our algorithm nearly matches the information-theoretically optimal exponential-time algorithm for the same problem due to Borgs et al. (FOCS 2018). More generally, we give an optimal, computationally efficient, private algorithm for estimating the edge-density of any graph whose degree distribution is concentrated in a small interval.
Researcher Affiliation Academia Adam Sealfon MIT and UC Berkeley asealfon@berkeley.edu Jonathan Ullman Northeastern University jullman@ccs.neu.edu
Pseudocode Yes Algorithm 1: Estimating the edge density of a concentrated-degree graph. Algorithm 2: Estimating the parameter of an Erd os-Rényi graph.
Open Source Code No The paper does not provide any links or explicit statements about the release of its source code.
Open Datasets No The paper describes theoretical models (Erd os-Rényi graphs) and algorithms, focusing on mathematical analysis and proofs. It does not conduct empirical experiments using specific datasets, thus no access information for a training dataset is provided.
Dataset Splits No The paper describes theoretical models and algorithms, not empirical experiments, thus no information on validation splits is relevant or provided.
Hardware Specification No The paper focuses on theoretical analysis and algorithm design, not empirical experiments requiring specific hardware specifications.
Software Dependencies No The paper focuses on theoretical analysis and algorithm design, not empirical experiments requiring specific software dependencies and versions.
Experiment Setup No The paper focuses on theoretical analysis and algorithm design, not empirical experiments requiring specific hyperparameter values or training configurations.